Posters, short powerpoint presentations over the ICU nurses’ station, screen savers and/or badges were used to increase awareness during daily practice. 5. Institutional safety climate: Creating an environment and the perceptions that facilitate awareness‐raising about patient safety issues while guaranteeing consideration of HH improvement as a high priority at all levels, including: - active participation at both the institutional and individual levels; - awareness of individual and institutional capacity to change and improve (self‐ efficacy); and - partnership with patients and patient organizations. Creating involvement with superiors was challenging in some centres, depending highly on individuals. Partnership with patients and patient organizations was not addressed in this ICU‐based study. For all of these five key components the WHO has developed freely accessible background information, an implementation guide and tools such as power‐point presentations and instruction videos (http://www.who.int/gpsc/5may/tools/en/). Three to six months before the start of the intervention, study nurses and involved physicians attended a two‐day PROHIBIT workshop for training best practices and implementation science, where the implementation guide and the tools were introduced and discussed. Hospitals were encouraged to adapt the intervention program and the tools to their local context. The Central venous catheter intervention (CVCi) In the Hospitals of Geneva, in 2007, existing protocols related to CVC insertion and care were reviewed and updated by an interdisciplinary study group, which included members from anaesthesiology, infection control, and the nursing department. A detailed insertion checklist was defined by the study group based on evidence in the literature and by repeated practice testing in daily routine. The complete insertion procedure from patient preparation until dressing application was filmed for training purposes. For catheter care, a modular e‐learning programme was developed, including assistance with CVC insertion, infusate preparation, CVC manipulation, dressing change, CVC removal, and clinical surveillance and documentation (www.carepractice.net). This modular e‐learning programme was made available in English for the PROHIBIT study participants and is freely available. All modules feature detailed procedure sequences and are animated by short movies, icons, and keywords. Three to six months before the start of the intervention, study nurses and involved physicians attended a two‐day PROHIBIT workshop for training best practices and implementation science, where present local CVC insertion practices and optimal CVC insertion procedure were discussed. The e‐learning programme was introduced and attendants practiced CVC insertion in an intensive care unit‐ based skills lab. Hospitals were encouraged to adopt CVC insertion kits and line carts, where not already in use. Table 1B specifies the activities of the individual hospitals. Methods and Results Competing events approach Different ‘events’ could happen to patients with a CVC, apart from acquiring a CRBSI. Patients could not be in need of a CVC anymore, they could be discharged to another hospital with their CVC or they could die (without a CRBSI) in the ICU. As these events can preclude the occurrence of our outcome of interest, in this case CRBSI, they are called competing events. Additionally it is possible that not only our outcome of interest is affected but (one of) the other outcome(s) as well, thus possibly leading to changes in the incidence of the outcome of interest.[2‐6] Standard Kaplan Meier curves and hazard ratios do not account for the effects of competing events, thus leading to incorrect hazards. Adjusted hazard ratios can be estimated using the sub‐ distribution approach, where the censoring time of the competing events is prolonged to the longest time at risk resulting in the outcome of interest (CRBSI).[7, 8] In our data the longest time at risk resulting in an infection was 73 days. We used the following model in SAS (with CVC duration (CVCtime) set to 73 days for records without infection):
RkJQdWJsaXNoZXIy MTk4NDMw